This invention relates generally to conducting network-based authentication transactions, and more particularly, to methods and systems for enhancing the accuracy performance of authentication systems that conduct network-based authentication transactions.
The identity of a user may be verified using biometric data captured during network-based authentication transactions. Authentication systems that conduct network-based authentication transactions based on voice biometric data typically require users to recite a phrase which is used to verify the identity of the user. Such authentication systems have been known to prompt all individuals to recite the same phrase. However, imposters have been known to surreptitiously obtain voice biometric data of users and to impersonate users during remotely conducted network-based biometric authentication transactions by providing a false claim of identity supported by fraudulent biometric data in an effort to deceive an authenticating entity into concluding that the imposter is the person they claim to be. Such impersonations are known as spoofing.
Impostors currently use many methods to obtain or create fraudulent biometric data of others that can be submitted during authentication transactions. For voice biometric data imposters have been known to eavesdrop on networks during legitimate network-based biometric authentication transactions to surreptitiously obtain genuine voice biometric data of users reciting the requested phrase. The imposters then use the obtained biometric data for playback during fraudulent network-based authentication transactions. Such fraudulent biometric data are known to be difficult to detect using known liveness detection methods. Verifying that the biometric data presented during a network-based authentication transaction conducted at a remote location is from a live person at the remote location, is known as liveness detection or anti-spoofing.
Methods of liveness detection have been known to use text independent voice recognition techniques in an effort to prevent imposters from conducting fraudulent transactions with surreptitiously obtained voice biometric data. However, authentication systems that authenticate based on text independent voice recognition techniques frequently generate inaccurate results and thus have unacceptably low biometric authentication accuracy rates of performance.